Gerine L posted on Wednesday, June 17, 2015 - 1:36 am
I am testing the following model:
!Stability lekaT4 on lekaT3 (b1); microT4 on microT3 (b2); aniceT4 on aniceT3 (b3);
!cros seffects lekaT4 on microT3 (b4); lekaT4 on aniceT3 (b5); microT4 on lekaT3 (b6); microT4 on aniceT3 (b7); aniceT4 on lekaT3 (b8); aniceT4 on microT3 (b9);
!correlations lekaT3 with microT3 (b10); lekaT4 with microT4 (b11); lekaT3 with aniceT3 (b12); lekaT4 with aniceT4 (b13); microT3 with aniceT3 (b14); microT4 with aniceT4 (b15);
It is a model in which 3 variables (leka, micro, anice) are measured across 2 timepoints (t3 T4).
I tested group differences (boy/girl). I found that some paths are not different between boys and girls, for instance b1 and b2. Is it a problem if I constrain paths for these effects to be equal, if I don't do anything for the means (i.e., these are still estimated separately for boys and girls)?